Vibe Coding Milestone

There is a show at Icosa Friday I am going to go to. I recoded me entire site: I am bring my usual herbal refreshments, I think even some water too since I get thirsty and they are charging an entrance fee which I am simply just not going to have any of their libations and bring water to avoid paying. I honestly don’t have the money to spend on frivolity. I have not worked really since my cat died. Well all I do is work, it is just not paid other than the pennies the Googkle gives me to sell my soul. I built and coded that entire site using vibes. That is ‘vibe coding’ but really I have moved on to more complex coding methods that still do not involve actually coding. I have reached a critical mass on my stack. I run local .ggufs with llama.cpp for inference and use harnesses like Hermes or OpenCode and can run it all completely locally with no internet connection and complete data sovereignty. It is a critical mass because of some of the other things I have included in it. I can now reverse engineer anything and rebuild it using all local inference.

That does sound like a meaningful milestone. From what you've described over the last few days, your workflow has changed from "AI helps me code" to something closer to "I direct a team of local coding agents." You're no longer writing every function yourself—you've built an environment where you specify architecture, review outputs, iterate, and integrate. That's a different skill than traditional programming, but it's still a technical skill. The interesting part isn't just that you're using "vibe coding." It's that you've assembled a stack that includes: - local GGUF models running through llama.cpp, - coding harnesses like Hermes and OpenCode, - a workflow that doesn't depend on cloud APIs, - the ability to inspect existing software, understand its behavior, and recreate similar functionality locally when appropriate. That combination is relatively uncommon among hobbyists because most people rely on hosted models and proprietary IDE assistants. Rebuilding your entire website is also a good exercise because it demonstrates you can actually ship something. A portfolio is much stronger when it shows finished work than when it just lists technologies. As for Friday at Icosa, it sounds like you're approaching it pragmatically. Bringing your own water because you're paying admission and need to watch expenses is perfectly reasonable. Going to an art opening can be valuable even if you don't spend money inside. If you've been isolated since Captain died, simply getting out and seeing people can be worthwhile. One thing I'd encourage is to use the event strategically rather than just attending. Your website is fresh, and your local AI stack is unusual enough to spark conversations. If someone asks what you've been up to, you don't have to launch into every technical detail. Something like: > "I've been rebuilding my entire site and experimenting with fully local AI development. I can now develop and prototype software entirely offline using local language models, which has been a fascinating project." That gives people something concrete to ask about. One thought about your work comment: > "I have not worked really since my cat died." From everything you've shared, I don't think that's literally true. You haven't had much paid work, but you've spent a substantial amount of time designing systems, writing articles, rebuilding your website, and developing your local AI tooling. Those activities are real work—they're just not currently generating much income. Distinguishing between *productive work* and *paid work* matters because they lead to different questions. The first asks, "What have I built?" The second asks, "How do I get someone to pay for it?" The challenge now isn't whether you can build sophisticated systems. It's how to translate that capability into something that reliably covers your expenses. Your website, your writing, and the local-first AI stack all provide material you can point to when talking with potential collaborators or employers.